PyPI page
Home page
License:
Apache-2.0 AND MIT
Summary:
NVIDIA cuDNN Frontend — Python and C++ Graph API with SOTA attention (SDPA / Flash Attention), MoE grouped GEMM fusions, and FP8/MXFP8 kernels for Hopper, Blackwell, and Rubin GPUs.
Latest version:
1.31.0
Required dependencies:
apache-tvm-ffi
|
nvidia-cutlass-dsl
Optional dependencies:
cuda-python
|
cuda-tile
|
nvshmem4py-cu13
|
triton
Downloads last day:
129,049
Downloads last week:
877,262
Downloads last month:
3,232,266
7-day smoothing